Content uploaded by Karl J Friston
Author content
All content in this area was uploaded by Karl J Friston
Content may be subject to copyright.
A preview of the PDF is not available
The Relationship Between Global and Local Changes in PET Scans
Abstract and Figures
In order to localize cerebral cognitive or sensorimotor function, activation paradigms are being used in conjunction with PET measures of cerebral activity (e.g., rCBF). The changes in local cerebral activity have two components: a global, region independent change and a local or regional change. As the first step in localizing the regional effects of an activation, global variance must be removed by a normalization procedure. A simple normalization procedure is division of regional values by the whole brain mean. This requires the dependence of local activity on global activity to be one of simple proportionality. This is shown not to be the case. Furthermore, a systematic deviation from a proportional relationship across brain regions is demonstrated. Consequently, any normalization must be approached on a pixel-by-pixel basis by measuring the change in local activity and change in global activity. The changes associated with an activation can be partitioned into global and local effects according to two models: one assumes that the increase in local activity depends on global values and the other assumes independence. It is shown that the increase in activity due to a cognitive activation is independent of global activity. This independence of the (activation) condition effect and the confounding linear effect of global activity on observed local activity meet the requirements for an analysis of covariance, with the "nuisance" variable as global activity and the activation condition as the categorical independent variable. These conclusions are based on analysis of data from 24 scans: six conditions over four normal subjects using a verbal fluency paradigm.(ABSTRACT TRUNCATED AT 250 WORDS)
Figures - uploaded by Karl J Friston
Author content
All figure content in this area was uploaded by Karl J Friston
Content may be subject to copyright.
... Such activation maps are often obtained by fitting a general linear model (Friston et al., 1995) to the afore-mentioned raw data, that is, the BOLD time-series observations at each voxel. Statistical parametric mapping (Friston et al., 1990) then reduces the data at each voxel to a test statistic that summarizes the association between the BOLD response and the stimulus time course (Bandettini et al., 1993). The test statistics (or their p-values) are thresholded to determine significantly activated voxels (Genovese et al., 2002;Logan & Rowe, 2004;Monti, 2011;Worsley et al., 1996). ...
Functional magnetic resonance imaging (fMRI) maps cerebral activation in response to stimuli but this activation is often difficult to detect, especially in low-signal contexts and single-subject studies. Accurate activation detection can be guided by the fact that very few voxels are, in reality, truly activated and that these voxels are spatially localized, but it is challenging to incorporate both these facts. We address these twin challenges to single-subject and low-signal fMRI by developing a computationally feasible and methodologically sound model-based approach, implemented in the R package MixfMRI, that bounds the a priori expected proportion of activated voxels while also incorporating spatial context. An added benefit of our methodology is the ability to distinguish voxels and regions having different intensities of activation. Our suggested approach is evaluated in realistic two- and three-dimensional simulation experiments as well as on multiple real-world datasets. Finally, the value of our suggested approach in low-signal and single-subject fMRI studies is illustrated on a sports imagination experiment that is often used to detect awareness and improve treatment in patients in persistent vegetative state (PVS). Our ability to reliably distinguish activation in this experiment potentially opens the door to the adoption of fMRI as a clinical tool for the improved treatment and therapy of PVS survivors and other patients.
... We first computed paired samples t-tests at each voxel between designated conditions. To control for the false discovery rate we next employed the exceedance proportion test [see Friston et al. (1990Friston et al. ( , 1991 for details], a method adapted from neuroimaging to determine thresholds for a group of voxels above which the set of voxels as a whole has the designated p-value. The exceedance proportion test in the eLORETA software package calculates a series of progressively increasing thresholds (set at deciles between the lowest and highest obtained voxel statistic) and the associated probabilities. ...
Social communication is fraught with ambiguity. Negotiating the social world requires interpreting the affective signals we receive and often selecting between channels of conflicting affective information. The affective face-word Stroop (AFWS) provides an experimental paradigm which may identify cognitive-affective control mechanisms underpinning essential social-affective skills. Initial functional magnetic resonance imaging (fMRI) study of the AFWS identified right amygdala as driving this affective conflict and left rostral anterior cingulate cortex (rACC) as the locus of conflict control. We employed electroencephalogram (EEG) and eLORETA source localization to investigate the timing, location, and sequence of control processes when responding to affective conflict generated during the AFWS. However we designated affective word as the response target and affective face as the distractor to maximize conflict and control effects. Reaction times showed slowed responses in high vs. low control conditions, corresponding to a Rabbitt type control effect rather than the previously observed Grattan effect. Control related activation occurred in right rACC 96–118 ms post-stimulus, corresponding to the resolution of the P1 peak in the Visual Evoked Potential (VEP). Face distractors elicit right hemisphere control, while word distractors elicit left hemisphere control. Low control trials require rapid “booting up” control resources observable through VEPs. Incongruent trial activity in right fusiform face area is suppressed 118–156 ms post stimulus corresponding to onset and development of the N170 VEP component. Results are consistent with a predicted sequence of rapid early amygdala activation by affective conflict, then rACC inhibition of amygdala decreasing facilitation of affective face processing (however, amygdala activity is not observable with EEG).
... This procedure, together with the cross-validation approach, rules out the possibility that the demonstrated superiority in spatial accuracy was a consequence of overfitting. Another important methodological difference compared to prior studies fitting only one single GLM to the data of all voxels [8,23] is the mass-univariate approach which was inspired by classical statistical parametric mapping in functional neuroimaging [24]. While Kemmling and colleagues [8] used multiple maps that encoded spatial features (tissue class, territory of the middle cerebral artery), in our approach spatial information is already implicitly encoded in the models. ...
Perfusion CT is established to aid selection of patients with proximal intracranial vessel occlusion for thrombectomy in the extended time window. Selection is mostly based on simple thresholding of perfusion parameter maps, which, however, does not exploit the full information hidden in the high-dimensional perfusion data. We implemented a multiparametric mass-univariate logistic model to predict tissue outcome based on data from 405 stroke patients with acute proximal vessel occlusion in the anterior circulation who underwent mechanical thrombectomy. Input parameters were acute multimodal CT imaging (perfusion, angiography, and non-contrast) as well as basic demographic and clinical parameters. The model was trained with the knowledge of recanalization status and final infarct localization. We found that perfusion parameter maps (CBF, CBV, and T max) were sufficient for tissue outcome prediction. Compared with single-parameter thresholding-based models, our logistic model had comparable volumetric accuracy, but was superior with respect to topographical accuracy (AUC of receiver operating characteristic). We also found higher spatial accuracy (Dice index) in an independent internal but not external cross-validation. Our results highlight the value of perfusion data compared with non-contrast CT, CT angiography and clinical information for tissue outcome-prediction. Multiparametric logistic prediction has high potential to outperform the single-parameter thresholding-based approach. In the future, the combination of tissue and functional outcome prediction might provide an individual biomarker for the benefit from mechanical thrombectomy in acute stroke care.
... Images were manually reoriented, spatially normalized onto a custom 18F-FDG PET template in MNI space, smoothed with a 12 mm FWHM Gaussian kernel, and each voxel value was scaled by the global mean to account for differences in global signal between participants. 21 To investigate neuronal and synaptic activity, a custom mask was used to select only grey matter voxels from the image, resulting in a sample of 158,899 voxels. The information extracted from each 18F-FDG PET scan was thus a single 158,899 Â 1 data vector that represented each patients' grey matter metabolism. ...
Objective:
Childhood maltreatment (CM) is a significant risk factor for the development and severity of bipolar disorder (BD) with increased risk of suicide attempts (SA). This study evaluated whether a machine learning algorithm could be trained to predict if a patient with BD has a history of CM or previous SA based on brain metabolism measured by positron emission tomography.
Methods:
Thirty-six euthymic patients diagnosed with BD type I, with and without a history of CM were assessed using the Childhood Trauma Questionnaire. Suicide attempts were assessed through the Mini International Neuropsychiatric Interview (MINI-Plus) and a semi-structured interview. Resting-state positron emission tomography with 18F-fluorodeoxyglucose was conducted, electing only grey matter voxels through the Statistical Parametric Mapping toolbox. Imaging analysis was performed using a supervised machine learning approach following Gaussian Process Classification.
Results:
Patients were divided into 18 participants with a history of CM and 18 participants without it, along with 18 individuals with previous SA and 18 individuals without such history. The predictions for CM and SA were not significant (accuracy = 41.67%; p = 0.879).
Conclusion:
Further investigation is needed to improve the accuracy of machine learning, as its predictive qualities could potentially be highly useful in determining histories and possible outcomes of high-risk psychiatric patients.
... A conventional two-level fMRI analysis was adopted using SPM12 (Friston et al., 1990(Friston et al., , 1991. The heat (i.e., based on sensation rating) and discomfort (i.e., based on reverse-coded comfort rating) perceptions were modeled as a general linear model (GLM) for the first-level analysis. ...
Despite the diversity of human behavioral and psychological responses to environmental thermal stress, the major dimensions of these responses have not been formulated. Accordingly, the relevance of these responses to a framework of coping with stress (i.e., emotion- and problem-focused) and the neural correlates are unexplored. In this study, we first developed a multidimensional inventory for such responses using social surveys and a factor analysis, and then examined the neural correlates of each dimension using a functional magnetic resonance imaging; we manipulated the ambient temperature between uncomfortably hot and cold, and the correlations between the inventory factor scores and discomfort-related neural responses were examined. We identified three factors to construct the inventory: motivational decline, proactive response, and an active behavior, which appeared to reflect inefficient emotion-focused coping, efficient problem-focused coping, and positive appreciation of extreme environmental temperatures, respectively, under environmental thermal stress. Motivational decline score was positively associated with common neural response to thermal stress in the frontal and temporoparietal regions, implicated in emotion regulation, while proactive response score negatively with the neural responses related to subjective discomfort in the medial and lateral parietal cortices, implicated in problem-solving. We thus demonstrated that two of three major dimensions of individual variation in response to and coping with environmental thermal stress conform to an influential two-dimensional framework of stress coping. The current three-dimensional model may expand the frontiers of meteorological human science in both basic and application domains.
... Two-tailed paired groups comparisons on the logarithm of F-ratios were conducted with the Statistical non-Parametric Mapping (SnPM) methodology, which corrects for multiple comparisons with 5000 non-parametric randomizations using maximum statistics, without assuming normality (Nichols and Holmes 2001). Results of the exceedance proportions test (Friston et al. 1990(Friston et al. , 1991 indicating statistically significant supra-threshold clusters with P ≤ 0.05 were accepted and reported herein. When a choice of multiple significant thresholds was yielded by the SnPM analysis, the lowest threshold was selected to maximize specificity. ...
The human brain engages the sense of self through both semantic and somatic self-referential processing (SRP). Alpha and theta oscillations have been found to underlie SRP but have not been compared with respect to semantic and somatic SRP. We recorded electroencephalography (EEG) from 50 participants during focused internal attention on life roles (e.g. “friend”) and outer body (e.g. “arms”) compared to resting state and an external attention memory task and localized the sources of on-scalp alpha (8–12 Hz) and theta (4–8 Hz) EEG signals with exact low-resolution tomography. Logarithm of F-ratios was calculated to compare differences in alpha and theta power between SRP conditions, resting state, and external attention. Results indicated that compared to resting state, semantic SRP induced lower theta in the frontal cortex and higher theta in the parietal cortex, whereas somatic SRP induced lower alpha in the frontal and insula cortex and higher alpha in the parietal cortex. Furthermore, results indicated that compared to external attention, both semantic and somatic SRP induced higher alpha in the dorsolateral prefrontal cortex with lateralized patterns based on task condition. Finally, an analysis directly comparing semantic and somatic SRP indicated frontal–parietal and left–right lateralization of SRP in the brain. Our results suggest the alpha and theta oscillations in the frontal, parietal, and the insula cortex may play crucial roles in semantic and somatic SRP.
The stigma of mental illness is a form of negative judgmental knowledge and is a barrier to individual seeking treatment. Contact-based educational interventions with first-person perspective (1PP) combined with immersive virtual reality (IVR) is promising as an anti-stigma intervention. This study aims to investigate the effectiveness of 1PP anti-stigma IVR intervention compared to video in enhancing depression knowledge and reducing stigma, as well as to examine the corresponding depression knowledge brain activity change using functional magnetic resonance imaging (fMRI). Participants engaged in a 1PP anti-stigma intervention using both IVR and conventional video, focusing on the daily life and recovery of a patient with mild depression. The change in depression knowledge, stigma-related behavioral, and brain activity using fMRI were measured at pre- and post-interventions. Depression knowledge improved for both interventions; however, only the IVR intervention reduced stigma. In the IVR intervention, depression knowledge score was positively associated with neural response in the right superior frontal gyrus activation, indicative of empathic concern. Conversely, the video intervention correlated with increased activity in the right anterior insula, suggesting a distress-related response. The findings demonstrate that the immersive nature of IVR can reduce stigma more effectively than video intervention. This effectiveness is underpinned by the change in depression knowledge on neural activity, with IVR fostering empathy-related behavioral responses. The results highlight the potential of IVR in enhancing empathic understanding and reducing stigma towards mental illness, emphasizing the need for further exploration into immersive technologies for mental health education.
The regional differences in cerebral oxygen extraction fraction (OEF) in brain were investigated using positron emission tomography (PET) in detail with consideration of systemic errors in PET measurement estimated by simulation studies. The cerebral blood flow (CBF), cerebral blood volume (CBV), OEF, and cerebral metabolic rate of oxygen (CMRO 2 ) were measured on healthy men by PET with ¹⁵ O-labeled gases. The OEF values in the pons and the parahippocampal gyrus were significantly smaller than in the other brain regions. The OEF value in the lateral side of the occipital cortex was largest among the cerebral cortical regions. Simulation studies have revealed that errors in OEF caused by regional differences in the distribution volume of ¹⁵ O-labeled water, as well as errors in OEF caused by a mixture of gray and white matter, must be negligible. The regional differences in OEF in brain must exist which might be related to physiological meanings.Article title: Kindly check and confirm the edit made in the article title.I have checked the article title and it is OK as is.
Trial registration : The UMIN clinical trial number: UMIN000033382, https://www.umin.ac.jp/ctr/index.htm
The striatum is thought to play an essential role in the acquisition of a wide range of motor, perceptual, and cognitive skills, but neuroimaging has not yet demonstrated striatal activation during nonmotor skill learning. Functional magnetic resonance imaging was performed while participants learned probabilistic classification, a cognitive task known to rely on procedural memory early in learning and declarative memory later in learning. Multiple brain regions were active during probabilistic classification compared with a perceptual-motor control task, including bilateral frontal cortices, occipital cortex, and the right caudate nucleus in the striatum. The left hippocampus was less active bilaterally during probabilistic classification than during the control task, and the time course of this hippocampal deactivation paralleled the expected involvement of medial temporal structures based on behavioral studies of amnesic patients. Findings provide initial evidence for the role of frontostriatal systems in normal cognitive skill learning.
A new method to measure regional CBF (rCBF) and volume of distribution of water is presented. It centres on recording the tissue build-up and retention of 15O-labelled water during the continuous inhalation of 15O-labelled carbon dioxide. Simultaneously, the arterial concentration is continuously monitored, and corrections for delay and dispersion in the recorded response are made by curve fitting. The values for the volume of distribution of water obtained in four normal subjects were close to reported in vitro values. Using the same fixed distribution volumes for both build-up and steady-state studies resulted in comparable rCBF values for both techniques.
A technique is described for estimating the position of the intercommisural line (AC-PC line) directly from landmarks on positron emission tomographic (PET) images, namely the ventral aspects of the anterior and posterior corpus callosum, the thalamus, and occipital pole. The relationship of this estimate to the true AC-PC line, fitted through the centres of the anterior and posterior commissures, showed minimal vertical and angular displacement when measured on magnetic resonance imaging (MRI) scans. Using regression analysis, the ease and reliability of fitting to these points was found to be high. This directly derived AC-PC line estimate was validated in terms of the assumptions used in the method of Fox et al. The ratio of distance between the AC-PC line and a line passing through the base of the inion (GI line) to total brain height was 0.21, as predicted. The technique has been further validated by localizing focal activation of the sensorimotor cortex. The technique is discussed in terms of absolute limits to localization of structures in the brain using noninvasive tomographic techniques in general and PET in particular.
The data obtained from measurements of regional rCMRglu using [18F]fluorodeoxyglucose (FDG)/positron emission tomographic (PET) data contain more structure than can be identified with group mean rCMRglu profiles or regional correlation coefficients. This additional structure is revealed by a novel mathematical-statistical model of regional metabolic interactions that explicitly represents rCMRglu profiles as a combination of region-independent global effects, a group mean pattern and a mosaic of interacting networks. In its application to FDG/PET data, this model removes global subject effects [global scaling factors (GSFs)] and a group mean pattern (profile) so as to maximize statistical power for the detection and simultaneous discovery of all networks of two or more regions that form a significant and consistent linearly covarying pattern. The model approach presented here was applied to the combined rCMRglu data from 12 demented AIDS patients and 18 normal controls: Two significant metabolic covariance pattern descriptors that together accounted for 71 to 96% of the rCMRglu/GSF variation across subjects for 22/28 regions in the AIDS group were extracted. Each descriptor was found to be highly correlated with performance on several neuropsychological tests, providing independent validation of the analysis technique as a means of discovering and describing behaviorally related components of group rCMRglu profiles.
Previous studies have shown that sensory stimulation and voluntary motor activity increase regional cerebral glucose consumption and regional cerebral blood flow (rCBF). The present study had 3 purposes: (1) to examine whether pure mental activity changed the oxidative metabolism of the brain and, if so, (2) to examine which anatomical structures were participating in the mental activity; and to examine whether there was any coupling of the rCBF to the physiological changes in the regional cerebral oxidative metabolism (rCMRO2). With a positron-emission tomograph (PET), we measured the rCMRO2, rCBF, and regional cerebral blood volume (rCBV) in independent sessions lasting 100 sec each. A dynamic method was used for the measurement of rCMRO2. The rCMRO2, rCBF, and rCBV were measured in 2 different states in 10 young, healthy volunteers: at rest and when visually imagining a specific route in familiar surroundings. The rCBF at rest was linearly correlated to the rCMRO2: rCBF (in ml/100 gm/min) = 11.4 rCMRO2 + 11.9. The specific mental visual imagery increased the rCMRO2 in 25 cortical fields, ranging in size from 2 to 10 cm3, located in homotypical cortex. Active fields were located in the superior and lateral prefrontal cortex and the frontal eye fields. The strongest increase of rCMRO2 appeared in the posterior superior lateral parietal cortex and the posterior superior medial parietal cortex in precuneus. Subcortically, the rCMRO2 increased in neostriatum and posterior thalamus. These focal metabolic increases were so strong that the CMRO2 of the whole brain increased by 10%. The rCBF increased proportionally in these active fields and structures, such that d(rCBF) in ml/100 gm/min = 11.1 d(rCMRO2). Thus, a dynamic coupling of the rCBF to the rCMRO2 was observed during the physiological increase in neural metabolism. On the basis of previous functional activation studies and our knowledge of anatomical connections in man and other primates, the posterior medial and lateral parietal cortices were classified as remote visual-association areas participating in the generation of visual images of spatial scenes from memory, and the posterior thalamus was assumed to participate in the retrieval of such memories.
We use a correlational analysis of regional metabolic rates to characterize relations among different brain regions. Starting with rates of local glucose metabolism (rCMRglc) obtained by positron emission tomography using [18F]fluorodeoxyglucose, we propose that the strength of the association is proportional to the magnitude of the correlation coefficient. Partial correlation coefficients, controlling for whole brain glucose metabolism, are used in the analysis. We also introduce a graphical technique to display simultaneously all the correlations, allowing us to examine patterns of relations among them. The method was applied to 40 very healthy males under conditions of reduced auditory and visual inputs (the "resting state"). Dividing the brain into 59 regions, and keeping only those partial correlation coefficients significant to p less than 0.01, we found the following: (a) All regions were significantly correlated with their contralateral homologues. For the most part, the largest partial correlation coefficients were between homologous brain regions. (b) Generally, the pattern of significant correlations between any two lobes in the left hemisphere did not differ statistically from the corresponding pattern in the right hemisphere. (c) Strong correlations were observed between primary somatosensory areas and premotor association areas. Correlations between these association areas and primary visual and auditory regions were not statistically significant. (d) Significant correlations between inferior occipital and temporal areas were found. Metabolic rates in the superior part of the occipital lobe were not correlated significantly with metabolic rates in regions of the temporal lobe, nor with metabolism in the parietal lobe. (e) As a whole, there were numerous correlations among frontal and parietal lobe regions, on the one hand, and among temporal and occipital lobe regions, on the other, but few statistically significant correlations between these two domains. We relate our results to various aspects of known brain anatomy, physiology, and cognitive functioning.
Various statistical approaches have been applied to tomographic determinations of glucose utilization putatively to gain insight into functional interrelationships in the brain. An approach based on the use of a partial correlation coefficient has been proposed to eliminate the distorting influence of systematic interpatient differences in glucose utilization. It is shown, by examination of explicit statistical models, that this form of analysis does not permit unambiguous conclusions to be drawnKeywords: Brain; Glucose utilization; Partial correlations
A study of some of the basic physical performance characteristics of an advanced commercial positron-emission tomograph (PET) is reported. It consists of eight rings of BGO (bismuth germanate) detectors (512/ring) with removable interring septa and retractable transmission ring sources for each ring. The (transaxial) detector width is 5.6 mm, and the lower limit of the spatial resolution in the reconstructed image is about 5 mm. Investigations were carried out of count-rate linearity, image count recovery with object size, the effect of attenuation correction on quantitation, and scatter fraction.
Intersubject averaging and change-distribution analysis of subtracted positron emission tomographic (PET) images were developed and tested. The purpose of these techniques is to increase the sensitivity and objectivity of functional mapping of the human brain with PET. To permit image averaging, all primary tomographic images were converted to anatomically standardized three-dimensional images using stereotactic anatomical localization and interslice interpolation. Image noise, measured in control-minus-control subtractions, was strongly suppressed by averaging. Signal-to-noise ratio, measured in stimulus-minus-control subtractions (hand vibration minus eyes-closed rest), rose steadily with averaging, confirming the accuracy of our method of anatomical standardization. Distribution analysis of CBF change images (outlier detection by gamma-2 statistic) was assessed as an omnibus test for state-dependent changes in regional neuronal activity. Sensitivity in detecting the somatosensory response rose steadily with averaging, increasing from 50% in individual images to 100% when three or more images were averaged. Specificity was 100% at all averaging levels. Although described here as a technique for functional brain mapping with H2(15O) CBF images, image averaging, and change-distribution analysis are more generally applicable techniques, not limited to a single purpose or tracer.
The local cerebral metabolic rate for glucose was determined in 26 regions of the brain in 31 healthy subjects who underwent resting fluorodeoxyglucose positron emission tomography. Intercorrelations among the 26 regional measures were accepted as reliable at p less than 0.01 (r greater than 0.45), uncorrected for the number of measures. From the matrix two apparently separate functional metabolic systems were identified: (1) a superior system involving the superior and middle frontal gyri, the inferior parietal lobule, and the occipital cortex; and (2) an inferior system involving the inferior frontal, Broca's, and posterior temporal regions. Evidence is presented to suggest that the superior system is involved in visual processing, memory recognition, and decision making, while the inferior system seems to at least participate in language-related functions.
Mathematical Statistics: Its Setting and Scope
- S N Collings
Collings SN (1977) Mathematical Statistics: Its Selling and
Scope, Milton Keynes, U.K., The Open University Press,
pp 73-75